All-To-All Broadcast and Matrix Multiplication in Faulty SIMD Hypercubes

In this paper, we develop algorithms in order of efficiency for all-to-all broadcast problem in an N=2/sup n/-node n-dimensional faulty SIMD hypercube, Q/sub n/, with up to n-1 node faults. The algorithms use a property of a certain ordering of dimensions. Our analysis includes startup time (/spl alpha/) and transfer time (/spl beta/). We have established the lower bound for such an algorithm to be n/spl alpha/+(2N-3)L/spl beta/ in a faulty hypercube with at most n-1 faults (each node has a value of L bytes). Our best algorithm requires 2n/spl alpha/+2NL/spl beta/ and is near-optimal. We develop an optimal algorithm for matrix multiplication in a faulty hypercube using all-to-all broadcast and compare the efficiency of all-to-all broadcast approach with broadcast approach and global sum approach for matrix multiplication. The algorithms are congestion-free and applicable in the context of available hypercube machines.

[1]  Cauligi S. Raghavendra,et al.  Total exchange in faulty SIMD hypercubes , 1994, Proceedings of 8th International Parallel Processing Symposium.

[2]  George Karypis,et al.  Introduction to Parallel Computing , 1994 .

[3]  Cauligi S. Raghavendra,et al.  Broadcasting algorithms in faulty SIMD hypercubes , 1992, [1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing.

[4]  M. Heath,et al.  Matrix factorization on a hypercube multiprocessor , 1985 .

[5]  Parameswaran Ramanathan,et al.  Reliable Broadcast in Hypercube Multicomputers , 1988, IEEE Trans. Computers.

[6]  M. H. Schultz,et al.  Topological properties of hypercubes , 1988, IEEE Trans. Computers.

[7]  P. Sadayappan,et al.  Iterative Algorithms for Solution of Large Sparse Systems of Linear Equations on Hypercubes , 1988, IEEE Trans. Computers.

[8]  Cauligi S. Raghavendra,et al.  Prefix Computation on a Faulty Hypercube , 1993, 1993 International Conference on Parallel Processing - ICPP'93.

[9]  Ming-Yang Kao,et al.  Optimal Broadcast in All-Port Wormhole-Routed Hypercubes , 1994, IEEE Trans. Parallel Distributed Syst..

[10]  Bülent Abali,et al.  Balanced Parallel Sort on Hypercube Multiprocessors , 1993, IEEE Trans. Parallel Distributed Syst..

[11]  Pierre Fraigniaud,et al.  Asymptotically Optimal Broadcasting and Gossiping in Faulty Hypercube Multicomputers , 1992, IEEE Trans. Computers.

[12]  S. F. Nugent,et al.  The iPSC/2 direct-connect communications technology , 1988, C3P.

[13]  S. Dutt,et al.  Parallel A* algorithms and their performance on hypercube multiprocessors , 1993, [1993] Proceedings Seventh International Parallel Processing Symposium.

[14]  Frank Thomson Leighton,et al.  Fast computation using faulty hypercubes , 1989, STOC '89.

[15]  Cauligi S. Raghavendra,et al.  Free dimensions-an effective approach to achieving fault tolerance in hypercube , 1992, [1992] Digest of Papers. FTCS-22: The Twenty-Second International Symposium on Fault-Tolerant Computing.

[16]  S. Lennart Johnsson,et al.  Optimum Broadcasting and Personalized Communication in Hypercubes , 1989, IEEE Trans. Computers.

[17]  Magdy A. Bayoumi,et al.  Efficient Mapping of ANNs on Hypercube Massively Parallel Machines , 1995, IEEE Trans. Computers.

[18]  Michael J. Quinn,et al.  Analysis and Implementation of Branch-and Bound Algorithms on a Hypercube Multicomputer , 1990, IEEE Trans. Computers.

[19]  John N. Tsitsiklis,et al.  Optimal Communication Algorithms for Hypercubes , 1991, J. Parallel Distributed Comput..

[20]  Ching-Tien Ho,et al.  Matrix Multiplication on Hypercubes Using Full Bandwith and Constant Storage , 1991, The Sixth Distributed Memory Computing Conference, 1991. Proceedings.

[21]  W. Daniel Hillis,et al.  The connection machine , 1985 .

[22]  Viktor K. Prasanna,et al.  Efficient Histogramming on Hypercube SIMD Machines , 1990, Comput. Vis. Graph. Image Process..

[23]  Bernd Becker,et al.  How robust is the n-cube? , 1986, 27th Annual Symposium on Foundations of Computer Science (sfcs 1986).

[24]  Cauligi S. Raghavendra,et al.  On some global operations in faulty SIMD hypercubes , 1996, Proceedings of International Conference on Parallel Processing.

[25]  Cauligi S. Raghavendra,et al.  Global semigroup operations in faulty SIMD hypercubes , 1993, [1993] Proceedings Seventh International Parallel Processing Symposium.

[26]  Cauligi S. Raghavendra,et al.  Global Commutative and Associative Reduction Operations in Faulty SIMD Hypercubes , 1996, IEEE Trans. Computers.

[27]  Jehoshua Bruck,et al.  Tolerating Faults in Hypercubes Using Subcube Partitioning , 1992, IEEE Trans. Computers.

[28]  Sartaj Sahni,et al.  Hypercube Algorithms: with Applications to Image Processing and Pattern Recognition , 1990 .

[29]  Fikret Erçal,et al.  Time-Efficient Maze Routing Algorithms on Reconfigurable Mesh Architectures , 1997, J. Parallel Distributed Comput..

[30]  Oscar H. Ibarra,et al.  Parallel Regognition and Parsing on the Hypercube , 1991, IEEE Trans. Computers.

[31]  Cauligi S. Raghavendra,et al.  Free Dimensions-An Effective Approach to Achieving Fault Tolerance in Hypercubes , 1995, IEEE Trans. Computers.